P3V3 is illuminated when RTM is inserted but disabled (hotswap)

[marmeladapk]

Problems

• When RTM power is disabled, FPGA unprogrammed
  • PDONE LED is on, however when I touch the gate of transistor with oscilloscope probe it turns off, I measure around 450 mV
  • P3V3 has 2,7V

Signals between AMC and RTM FPGAs and ICs

• RTM UART - 400 mV
• RTM_FPGA_CFG
  • INIT_B 1,4V
  • Done 400mV (but LED shuts off so measurement changes state)
  • Program_b 400mV
  • CCLK 400mV
  • DIN 400mV
• RTM I2C
  • 3,3V and talks (however it's connected to 3V3MP on AMC)
  • TMP_SENS is pulled up to P3V3 and has 2,8V (but it could be from another source, MMC doesn't seem to be talking there)
• RTM_JTAG - potential cause, has pullups, however without USB plugged in 3V3 on RTM is also high
  • TMS 3,3V
  • TDI 3,3V
  • TDO 1,7V
  • TCK 3,3V
  • AMC: PROG_TCK 3,2V
  • AMC has internal resistive pullup on these pins
• FPGA_LINK GTP

Fixes for ARTIQ

• rtm_fpga_cfg -> done pin - shouldn't have pullup - DONE signal already has internal pullup in RTM FPGA and sets signal high before RTM FPGA has power
• init_b, program_B shouldn't have pullups as well

Removing IC15 & IC13

• When I removed IC15 P3V3 voltage dropped to 2,5V from 2,7V when RTM was not powered
• When I removed IC13 & IC15 P3V3 voltage dropped to 0V + PDONE is no longer illuminated
• Only P3V3MP LED is illuminated
• Pressing hotswap turns on all supplies as expected

So IC15 and IC13 shouldn't supply high level signal to RTM when RTM is disabled.

ed: IC14 instead of IC15

[marmeladapk]

Possible ways to move forward with current boards:

• abandon hotswap on RTM until v2.1
• rework AMC boards
  • remove IC13, IC14
  • populate R59-R62 (JTAG bypass)
  • this will remove RTM and FMC from JTAG chain

Thoughts @jbqubit, @gkasprow ?

[gkasprow]

We don't need hot-swap anyway, do we? Software does not support it.

[jbqubit]

RTM hotswap is a low-priority feature. If skipping hotswap for now helps things move forward faster let's do that.